# NFB with Cathodyne

Discussion in 'Amp Tech Center' started by _Steve, Jul 29, 2020.

1. ### Mexitele BluesTele-Afflicted

Age:
42
Posts:
1,060
Joined:
Mar 7, 2018
Location:
Westminster, CO
I'm trying to grok what's happening in that nfb loop. NFB signal hits r10 and r35 in parallel with r34, approx 20% of the signal takes the lower path, across an rc low pass (r10 and c8) at 50hz and another at about 736hz (r35 and tail capacitor c23). Meanwhile the bulk of the signal takes the high road through c15, effectively a bright cap bypassing the rc filters below it, but then gets chopped in half by the capacitive voltage divider created by c15 and c23. The remaining nfb signal is then injected into the cathodyne grid through coupling cap c20.

Is any of that close to what's happening there?

Posts:
36,021
Joined:
Mar 17, 2003
Location:
Lubbock, TX
looking at the 5E11, we see a ‘top’ resistor in that NFB loop with a value of 56k. This is Rt. The cathode resistor into which that Rt feeds is a 1500ohm resistor...Ri. The formula for the negative feedback factor is Ri/(Ry+Ri).
1500/57500= .026

Many BF amps use 820ohms in the Rf and. 100 ohm in the Ri. 100/920 = .109. the BF circuit has more cancellation effect.

I have no idea what is going on with that PV circuit.

3. ### _SteveTDPRI Member

Age:
40
Posts:
20
Joined:
Jul 14, 2020
Location:
Los Angeles
Just an update that I got the amp finished and working today, which also happens to be our 10y wedding anniversary so im gonna call it the "Anniversary Special"

I'm sure i'll need to spend a bit of time over next few days dialling it in then i'll report back the sound and ask for ideas where to go from here. I'll also try to get both methods of NFB working and maybe record some clips.

Steve

(obviously still needs a lot of work)

Wally and elpico like this.
4. ### printer2Poster Extraordinaire

Posts:
6,740
Joined:
May 24, 2010
Location:
The lower part of the network is a T filter, it is a notch filter. As it is a notch it reduces the NFB in the notch area. Basically it is a resonance control. It simulates the impedance peak of the speaker resonance or a peak in the 2-3k band. There are equations to calculate the notch frequency but they screw it up by not using equal value. Had an idea. I used the online Duncan Tonestack Calculator and inserted values into the E-Series simulation.

RIN Ω 10 Output transformer impedance, just a guess
R2 Ω 22000000 High value to get it out of the equation
R3 Ω 470000 R10 in schematic
RB Ω 10000000 High value to get it out
R5 Ω 2200000 R35 in Sch
RT Ω 750000 R24 and volume pot,
C1 pF 100 C23 in sch
C2 pF1 Low value to get rid of it
C3 pF 100000 High value to get rid of it
C4 pF 6800 C8

And hit apply. Well that looks interesting, a gentle bump in the high end. But wait, have to turn up RB to its max and turn RT its minimum to hit my assumptions.

SOB. I had to walk away to laugh. A notch at 2 kHz. Those clever Peavey people. That was really cool. After looking at it for a second I realized I could get the rest of the filter network in it. Peavey C15 180 pF, I can use C1 of the calculator for it. And if I set RT just right I can use it for R34 470k. Just for the heck of it (I don't know if the pot is linear or log, have TSC on my other computer but I don't feel like firing it up) I will use RT as 1.5 M and split it into 1/3 and 2/3 so I get 500k and 1M as the 220k and volume control.

Hmmm.... ...not exactly what I thought I would see. It turned into a simple high pass filter at 2 kHz. Wait a second, I have the RT pot reversed, I should have it at 75%. No, that just cause a slight depression at 100 Hz. Moving the pot more to the right, all the way up.

Laugh again. Although it has a resonance at 140 Hz it is the mirror image of a speaker impedance plot. A low frequency resonance and a rising impedance starting at 1 kHz. Now turning the pot to its 9.9 value, it lowers the impedance peak and moves it to 120 Hz. Keep in mind that the TSC model was not designed for this messy circuit and it may be off a bit but I would put my money on the guys at Peavey putting in a speaker simulation curve (negative due to the NFB) to make the installed speaker and amp combination sound more 'tubey'.

And you wonder why I don't get anything done all day.

5. ### _SteveTDPRI Member

Age:
40
Posts:
20
Joined:
Jul 14, 2020
Location:
Los Angeles
Wow that is an insane amount of investigation and insight! Of the tiny portions I understand, I am impressed

It's a pretty intriguing concept - i wonder if it worked well?

6. ### tubegeekFriend of Leo's

Age:
59
Posts:
2,486
Joined:
Jan 31, 2020
Location:
Brooklyn, NY
Naah. I don't.

7. ### David BarnettDoctor of Teleocity

Age:
64
Posts:
12,973
Joined:
Mar 17, 2003
Location:
The Far-Flung Isles of Langerhans
I would argue that's not the final preamp tube, but the front end of the power amp.

robrob likes this.

Posts:
36,021
Joined:
Mar 17, 2003
Location:
Lubbock, TX
Is not that triode of V4 both a final preamp gain stage and a driver for the cathodyne circuit that operates in the other triode of V4?
It helps make up for gain lost in the reverb circuit while also giving that less-than-unity gain of the cathodyne a bit of a push, doesn’t it?

9. ### Ten OverTele-Holic

Posts:
605
Joined:
May 13, 2015
Location:
Central California
I'm calling the 56K resistor in the 5E11 NFB the "feedback resistor" and designate it Rf. I am designating the 1.5K cathode resistor Rk. Rt is the "tail resistor" and it equals Rk in parallel with the internal cathode resistance. The feedback factor is Rt/(Rt+Rf) and I am calling it "Beta". If the internal cathode resistance is 1.5K, then Rt=750 and Beta is 0.013

The ratio of the closed-loop gain to the open-loop gain is what we are really interested in and just Beta by itself doesn't give you enough data to compare this ratio between amps. The 820/100 NFB loop in some BF amps could well have more cancellation effect (less closed-loop gain to open-loop gain), but you can't tell this just by comparing Betas.

robrob likes this.
10. ### David BarnettDoctor of Teleocity

Age:
64
Posts:
12,973
Joined:
Mar 17, 2003
Location:
The Far-Flung Isles of Langerhans
But the same circuit exists in a non-reverb Princeton-Amp:

But yeah it's splitting semantic hairs, since a Princeton is an integrated guitar amp, not a set of hifi separates.

tubegeek likes this.
11. ### _SteveTDPRI Member

Age:
40
Posts:
20
Joined:
Jul 14, 2020
Location:
Los Angeles
Since there are some very smart people in here (and I have to get by mainly on my looks), can someone please explain what the effect (if any) of NFB is on the OT primary impedance as seen by the power tubes? Assuming that the NFB is connected to the same tap as the speaker..

tubegeek likes this.
12. ### tubegeekFriend of Leo's

Age:
59
Posts:
2,486
Joined:
Jan 31, 2020
Location:
Brooklyn, NY
That's a great question. If nobody comes up with a good explanation I will hit the books and see what I can find. My gut feeling is: no change, but it will increase the damping factor of the amp/speaker combination. But I honestly am not sure at all..

13. ### printer2Poster Extraordinaire

Posts:
6,740
Joined:
May 24, 2010
Location:
I saw the schematic before, had a rough idea what it was doing but I never had reason to consider using NFB in the PI. I did want to use it with a Mosfet in the place of a triode in the cathodyne but never got around to it. Now I think I will give it a shot in an amp.

Did it work? Peavey thought so

Should not matter. But thinking about it the NFB is suppose to lower the output impedance of the amp. As a quick thought, the NFB takes in the back EMF from the speaker and any error, namely the resonance, the bigger the back EMF error. This error gets inverted and makes its way through the amp and corrects the speaker motion an the output impedance appears to be lower.

Last edited: Aug 1, 2020

Posts:
36,021
Joined:
Mar 17, 2003
Location:
Lubbock, TX
in that non-verb Princeton, we see that there is an input gain stage for each channel followed by that second gain stage/cathodyne PI driver stage triode. It is a 1+1 gain structure prior to that cathodyne PI triode as I see it. Maybe not??

15. ### _SteveTDPRI Member

Age:
40
Posts:
20
Joined:
Jul 14, 2020
Location:
Los Angeles
So some feedback for those interested...

I've played the amp a bunch today and messed around a bit with the NFB.

The clean tone is quite good! Not exceptional, but workable if I drive mainly with Normal and blending in some of the Bright channel like you would a Bassman/Plexi. Maybe a little thin? But overall better than I was expecting.

Personally I think it sounds best with no NFB and no bypass cap on V2a. Without that bypass cap it never gets woolly and theres just no need for the NFB, which just dulls the whole thing.

The Tremolo is a blast! OK its my first ever tremolo amp but it sounds super lush. Its a little on the speedy side so Im going to slow it down a tad.

HERES THE SURPRISE:

Just for kicks I cranked everything. Just below maxed out this thing is a crunchy, punchy, no-fizz rock machine! Like seriously very impressive for a Marshall guy like me. Not at all loose, farty or fuzzy like I expected, especially with the Cathodyne. I guess if I wanted to go high-gain (would need reconnecting V2a bypass cap) that might become an issue. There is a very, very slight ringing or harmonic or something that I can hear on certain notes which i guess could be related to the cathodyne, im not sure.

Looks like I have a knack for achieving the complete opposite of what I set out to do LOL. To get a good clean-sounding amp looks like going to have to try to build a Soldano or something....

Last edited: Aug 3, 2020 at 7:28 PM
robrob and hepular like this.
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.